Bài kiểm tra có liên kết hướng dẫn · Trung bình Cấp độ

AI trong bài kiểm tra phân tích xét nghiệm máu và huyết học

Kiểm tra hình thái kỹ thuật số, phân tích hình ảnh bệnh sốt rét, xác thực trên các slide và sự khác biệt giữa cờ tế bào và chẩn đoán.

Đường dẫn hướng dẫn liên quanAI trong phân tích huyết học phết máu
Câu hỏi 1 của 8

What can an AI digital-morphology tool do with a blood-smear image?

Xem trước mọi câu hỏi8 câu hỏi

Câu trả lời và giải thích xuất hiện khi bạn làm bài kiểm tra.

  1. What can an AI digital-morphology tool do with a blood-smear image?
    • Estimate kidney risk from serial creatinine results
    • Classify or flag cells for trained review
    • Score fibrosis on a liver-biopsy slide
    • Predict cancer origin from plasma DNA
  2. How many patients were included in the AIDMAN prospective smear validation?
    • 64 image fields from one test slide
    • 640 patients across multiple hospitals
    • 64 hematology laboratories
    • 64 patients with thin blood smears
  3. What information helps interpret an unusual cell flag?
    • A ranked list of model labels without checking the slide
    • The model confidence value without other laboratory results
    • The original smear, blood count and clinical context
    • A training-set accuracy figure without patient-level context
  4. Which source can contribute to image-model performance drift?
    • Different stains, scanners or magnifications
    • The use of an electronic lab order
    • A clinician correcting a flag
    • A cell being labeled on screen
  5. Which follow-up can further characterize a suspected hematologic malignancy beyond cell morphology?
    • A second image classifier on the same smear only
    • Flow cytometry or molecular testing
    • A symptom questionnaire by itself
    • A routine cell count without morphology review
  6. Why is a high cell-level score not proof that a system can rule out disease?
    • A high cell score proves the model works in every lab
    • An image-level measure alone shows the test rules out disease
    • Model confidence replaces a clinical reference standard
    • Cell classification does not automatically establish patient-level performance
  7. What should a useful laboratory workflow preserve?
    • Automatic acceptance of the first classification
    • Using AI output as the final treatment decision
    • Access to original images and professional correction of labels
    • Removing manual review from the workflow
  8. What does a cell-level AI flag establish by itself?
    • Only that the software marked an image for attention
    • A confirmed disease diagnosis
    • A required medication
    • That additional testing is unnecessary